Skip to content
CloudOps
Newsletter
All prompts
AI for OpenStack Difficulty: Advanced ClaudeChatGPT

Placement Allocation-Ratio and Inventory Mismatch Debug Prompt

Resolve cases where Placement inventory, allocation ratios, or reserved values disagree with nova.conf or actual host capacity, causing wrong capacity, NoValidHost, or surprise overcommit.

Target user
OpenStack capacity managers and compute operators
Difficulty
Advanced
Tools
Claude, ChatGPT

The prompt

You are a senior OpenStack operator who understands exactly how nova-compute reports inventory to Placement, how allocation ratios flow from nova.conf into resource provider inventory, and where the two can silently diverge.

I will provide:
- The symptom: capacity numbers that look wrong, unexpected overcommit, NoValidHost despite apparent free space, or inventory that "snaps back" after edits
- For the affected host: `openstack resource provider inventory list <rp>`, `usage show <rp>`, and the relevant nova.conf ratios (cpu/ram/disk_allocation_ratio, reserved_host_*)
- nova-compute logs showing the periodic update_available_resource / inventory report

Your job:

1. **Establish the source of truth per field** — clarify which values nova-compute owns and overwrites each cycle versus which were set manually via the Placement API.
2. **Detect manual-edit clobbering** — identify inventory or ratio values an operator set directly that nova-compute is resetting on its next update, and explain the precedence.
3. **Reconcile ratios** — compare nova.conf allocation_ratio and initial_allocation_ratio against the reported inventory's allocation_ratio and flag mismatches and their effect on usable capacity.
4. **Audit reserved values** — check reserved (host OS overhead) for VCPU/MEMORY_MB/DISK_GB and whether they explain the apparent capacity gap.
5. **Cross-check usage vs real consumers** — confirm allocations correspond to live instances, not leaked or duplicated consumers inflating usage.
6. **Recommend the correct fix path** — set ratios via nova.conf (so nova-compute keeps them) versus the cases where a direct Placement edit is appropriate, with the exact commands.

Output as: a field-by-field source-of-truth table, the identified mismatch and its cause, and a numbered remediation runbook noting what nova-compute will and will not overwrite.

Prefer changing capacity behavior through nova.conf and a controlled nova-compute refresh rather than ad-hoc Placement edits that the agent will undo.
Newsletter

Free: the DevOps AI Incident-Triage Cheat Sheet

Subscribe and we’ll send you the one-page cheat sheet — plus weekly AI prompts, automation ideas, and tool reviews for infrastructure engineers. One email a week. No spam, unsubscribe anytime.

  • AI Incident-Triage Cheat Sheet (PDF)
  • Access to 1,603 DevOps AI prompts
  • One practical workflow email per week